• A translation which closely follows the original and is chiefly meant as an aid to understand the Sanskrit text (as in Kataoka 2011)
• A translation which smooths the text, so that it sounds as if it had been originally written in the target language

It seems that the choice of approaching more the one than the other partly depends also on one’s target language and on the cultural expectations linked to a certain literary civilisation, since readers of, e.g., English or German appear to have very different expectations concerning what counts as a good translation, with the former being much more positively impressed by texts which sound as if they had been composed in their own language, whereas the latter tend to expect from the translation that it transmits part of the flair of the original language, as if it could be partly transparent and let one glimpse the original through it (see Venuti 1995). The other element of the choice is the reader one has in mind. The first model presupposes a reader who knows Sanskrit fairly well and uses the translation only as an auxiliary, the second one assumes the reverse.

A third element worth considering regards the translator themselves. They need an extremely good command of English in order to translate in the second way (which is not my case). Moreover, they need to think of the duration of one’s translation. Each language rapidly evolves so that translating in a very idiomatic way runs the risk of rendering the text less understandable for non-native speakers or speakers who will live in a not so distant future. As a non-native speaker of English, I have, for instance, had problems deciphering the English idioms used in a translation by Anand Venkatkrishnan (in Venkatkrishnan 2015, see the discussion here: http://wp.me/p3YaBu-jz) and many readers of mine had asked me what “ones” in Edgerton 1929 or “ain’t” in other authors could mean. In other words, a strongly idiomatic translation is restricted —in at least some aspects— to native speakers of the present
and immediately following decade.

My translation of the SM aims at being close enough to the text to make it conceptually understandable to a public of Sanskritists as well as to scholars of the history of philosophy and theology. It does not reproduce the complexity of Vedānta Deśika’s prose, nor does it attempt at capturing the beauty of his verses. I tried to make the translation readable also to non-Sanskritists, while being at the same time extremely cautious in not over translating the Sanskrit text. As a rule of thumb, I used the concept of functional equivalence, which was used, e.g., by Raimon Panikkar in discussing comparative theology. Accordingly, I changed passive forms (which are the rule in scholarly Sanskrit) into active ones, since the latter are the rule in scholarly English. I also made pronouns and copulae explicit and avoided causative clauses when possible, since this is the rule in scholarly English. I also broke long sentences into shorter ones, since the length of an acceptable sentence varies massively between scholarly Sanskrit and scholarly English. Nonetheless, the resulting text will not be smooth, because the source text is extremely complex and causes real head-aches to its readers. Smoothing it completely would have meant unpacking all the points it presupposes and hints at.

Whichever type of translation one favours, translating a Sanskrit text always remains a difficult task. In my opinion, this is due especially to the fact that contemporary readers lack almost completely the background assumptions which are required in order to understand each instance of communication, including philosophical texts. While reading a text by Plato, a contemporary reader educated in Europe will automatically be able to identify most of the trees, places (e.g., the Piraeus, at the beginning of the Republic), social institutions (e.g., the role of theater, the names of gods and goddesses, the presence of slaves) and the other realia he mentions. Readers will not be scared by reading geographic or ethnic names (e.g., hoi Thrâkes, in Republic A 327), nor by proper names (e.g. Socrates or Glaucon, ibid.). This familiarity is also reflected in the fact that there are English (as well as French, German, Italian, Spanish…) versions of these terms. Even more, readers will likely be at least slightly familiar with many of Plato’s ideas, such as the maieutic method or the realm of ideas.

The situation is completely different in the case of Sanskrit texts, where Euro-American readers often need to deal with unfamiliar terms, contexts and ideas. Let me call the lack of familiarity with names, works, contexts, customs, etc., circumstantial unfamiliarity and the lack of familiarity with philosophical ideas philosophical unfamiliarity. The unfamiliarity with ideas might be something readers are willing to live with —after all, they started reading a book about an unfamiliar philosopher. What they are probably not prepared to have to overcome, is the additional effort required to just overcome the circumstantial unfamiliarity.

For this reason, one might decide to strongly alter the text, in order to substitute the background assumptions with ones more familiar to the contemporary reader. This substitution may regard minor details, e.g., the substitution of “Devadatta” with “John Smith” as the placeholder for whoever a person, or the inclusion of pronouns, copulae, punctuation and other elements which can be deduced out of the context or of the literary usage of scholastic Sanskrit. I have been generous in making implicit linguistic units explicit, but I did not dare translating proper names and terms referring to realia and to culturally specific elements into more familiar ones. This is because I want to create a reliable translation of the SM that is readable for at least some decades —also given the fact that I do not foresee new translations
of the SM being prepared in the short- and mid-term. Furthermore, I did not want to limit my assumed readership to European readers, and I am not sure that “John Smith” will remain the standard way of expressing the same thing as the Sanskrit Devadatta for all future and remote readers.

Coming to philosophical unfamiliarity, there are again two kinds of it. On the one hand, there is the unfamiliarity of the main thesis one is reading about, in the present case, the idea of aikaśāstrya. On the other hand, there is the unfamiliarity of other philosophical ideas being mentioned only in passing. In the former case I believe I can expect the reader to be patient enough to tolerate some Sanskrit words and some translations sounding alien,
and give themselves enough time to understand what is being discussed. In the latter case, by contrast, I would like to provide the reader with enough information to go forward with the text without having to engage deeply with each of the ideas and theories being mentioned in it without extensive explanations. This means, that in a chapter focusing on the instrument of knowledge called śabda, I might expect the reader to bear with my complicated translation of it as `linguistic communication as an instrument of knowledge’, whereas I will not hesitate to translate anumāna just as `inference’ and adding a word of caution in, e.g., a footnote only. Vice versa, I would translate śabda just as `testimony’ (with a word of caution in, e.g., a footnote) while translating the chapter on anumāna by a thinker of the Nyāya school.

If this principle is followed, a reader will receive a lot information about some key terms and only minimal information about less relevant ones. How can one organise this information so that a reader can find all the information they need to go forward with the main thesis while understanding enough of the other philosophical ideas mentioned?

For this purpose, one needs more than just a reader-friendly translation. One might decide among at least three options:

1. Adding extensive comments in footnotes or endnotes (as it has been done in Preisendanz 1994)
2. Adding the same comments within the text in separate paragraphs, perhaps in smaller font size (as it has been done in Taber 2005)
3. Adding the same comments in an extended introduction (as it has been done in Bilimoria 1988)

The choice partly depends on one’s target readers. Philologists are more likely to appreciate the first solution, whereas the latter two are more likely to appeal to a public of more general readers, who might be more interested in the philosophical content than in the text itself. In this book, I adopt the third model (as I did already in Freschi 2012), although I will recur to the first one whenever the text demands a punctual explanation of a specific point having little bearing with its major concerns.

A specific paragraph needs to be dedicated to the use of parentheses and brackets. I used parentheses:

1. to indicate which Sanskrit word I am translating in specific cases (i.e., while introducing for the first time the translation of a technical term, or whenever a term has to be understood in an unexpected way).
2. to insert short explanations which are needed to understand a specific point of the text and cannot therefore be postponed in an explanatory footnote nor advanced in an introductory study.

As for square-brackets, I used them to insert words which were not present in the Sanskrit original and which could not be directly inferred on its basis. In other words, I would not put “I” in “I am going” within brackets if this translated gacchāmi, since the first-person subject is obviously present in the verb form. I also did not put within brackets obvious complements, such as “sacrifice” while translating sacrifice’s names such as darśapūrṇamāsau or citrā as `the full- and new-moon sacrifices’ and `the citrā sacrifice’. By contrast, I used brackets to highlight for the reader that a certain concept is not actually found in the text, so as to make them aware of the fact that I am suggesting an interpretation of it. For instance, God’s attribute vimuktipriya (within the opening verses of the SĀṬ) literally means `who is fond of liberation’. I rendered it as `He who wants [people to achieve] liberation’.
I also use brackets to introduce identifications of speakers.

Comments from fellow translators and/or readers of philosophical texts from afar welcome!

(Cross posted on the Indian Philosophy Blog. Read there some very interesting comments).

Let me start by trying to achieve some clarity. John Taber, in his exemplary book on the chapter on perception of the Ślokavārttika speaks of a “semi-critical edition”, insofar as he did not look at new manuscripts, but improved the text of the editions by collating and comparing them among each other, with the commentaries and with the sources. The result is an appendix with suggested readings for the ŚV text. It is hard to call it a “critical” edition (because it lacks a manuscript basis) and the label “semi-critical” is also possibly misleading, since it seems to suggest that the text has the same basis of the one of a critical edition, but that the critical choices have not been completed. By hearing “semi-critical”, I would, accordingly, rather expect something between a collation and a critical edition, not a text based on existing editions and further improved as described before. Therefore, I would call these cases just “improved edition“, but “revised edition” would also work. Still, in cases such as the Ślokavārttika, one can reasonably attempt the reconstruction of a specific text as based on either Kumārila’s original intention (as reconstructed through his other works, his interactions with other thinkers and his commentators) or on the text as read by a given commentator.

The colleague also invited me to consider the case of texts for which a critical edition could be said to be impossible, such as Purāṇas. Now, let me repeat that an edition might have different purposes, and that the key is to be aware of what one wants to achieve. Aiming at the Urtext of a hymn transmitted orally in different regions and perhaps even languages might be out of place, but one might reasonably aim at reconstructing the text of the same hymn as it was read and transmitted in manuscripts in, say, 16th c. Karṇaṭaka, or as commented upon by a given scholar in 13th c. Gujarat. Alternatively, one might try to reconstruct the history of the transmission. Charles Li and others have been developing softwares which allow one to put one or the other manuscript collated as the main text and the others as alternative readings and to change the main text with just one click. In this way, it would be relatively easy to compare the version of the text transmitted in one or the other group of manuscripts.

In other words, let us not mix the intense work needed to prepare a critical editions with only low level textual criticism or with the myth of the reconstruction of an Urtext. This is just one possible approach to the redaction of a critical editions and can only be appropriate in specific cases (authorial texts for which a historical author can be individuated and manuscript material can be proven to depend from a single source etc.).

How would you call “improved editions”?

In Classical Indian philosophy, Naiyāyika authors should uphold the possibility of a non-committal understanding of sentence-meanings, since they are convinced that cognitions need to be proved to be valid in order to be such and that such validation comes from outside (in the case of testimony, typically out of the reliability of the speaker). Mīmāṃsā authors, by contrast, would claim that belief is withdrawn but that the default understanding of a sentence implies the belief that it states something true.

Now, the problem with the non-committal understanding is that it seemse to have been theorised (and called śābdabodha) as such only in Navya Nyāya. Taber (1996) discusses the possibility of such a hypothesis already in the 10th c. Jayanta, but his arguments are critically analysed in Graheli (2015, forthcoming in Kervan).

By chance I read of a similar discussion in the Port Royal Logique in this post, where the problem is also how to account for the distinction between understanding a sentence and believing that its content is the case. The post is highly recommended!

Are words an instrument of knowledge? And, if so, what sort of? Are they an instance of inference insofar as one infers the meaning on the basis of the words used? Or are they are an independent instrument of knowledge, since the connection between words and meanings is not of inferential nature?

Kumārila discusses the topic in the Śabdapariccheda ‘Chapter on Words’ (or ‘Chapter on Linguistic Communication’) of his Ślokavārttika (henceforth ŚV). The chapter is the first one focusing specifically on language of the ŚV and in this sense it needs to hint to various arguments which will be elaborated upon in various successive chapters. Accordingly, in the same chapter Kumārila addresses Sāṅkhya, Naiyāyika, other Mīmāṃsaka, and Vaiśeṣika and Buddhist opponents.

I have dealt with the first group of opponents (of the Sāṅkhya, Nyāya and Mīmāṃsā schools) in this post. Against the Buddhist opponents, Kumārila’s strategy is to first show that no inference (in the technical sense of the Sanskrit anumāna) can be formalised in the case of word and meaning. Instead, words are described as being an independent instrument of knowledge. Then, however, two dramatic turns occur:

1. Kumārila candidly admits that he has only provisionally accepted that the words are instruments of knowledge. Instead, words are in fact not an instrument of knowledge. Only sentences are. Words, rather, repeat what has been acquired through another instrument of knowledge or function as recollectors (smāraka).
2. Next, Kumārila even more candidly admits that words can indeed be an instance of inference. This is nonetheless no problem for the advocates of Linguistic Communication as an instrument of knowledge since sentences, not words are the vehicles for knowledge.

This is not the first case in which Kumārila is ready to abandon a thesis in order to save a more general point (see John Taber’s Kumārila On Perception regarding perception and Erich Frauwallner’s Bhāvanā und Vidhiḥ bei Maṇḍanamiśra regarding the role of root and verbal ending in communicating the bhāvanā).

Does it mean that Kumārila is often chiefly a polemist? Or that he focuses more on other issues (in this case, more on the authority of the Veda)?

More information on the workshop on the Śabdapariccheda out of which this post originated can be read here.
Here you can read my analysis of the first part of the chapter.

• A translation which closely follows the original and is chiefly meant as an aid to understand the Sanskrit text (as in Kataoka 2011)
• A translation which smooths the text, so that it sounds as if it had been originally written in the target language (Dominik Wujastyk’s and Ch. Ram-Prasad’s ones)

It seems that the choice of approaching more the one than the other partly depends also on one’s target language, since readers of, e.g., English or German have very different expectations concerning what counts as a good translation, with the former being much more positively impressed by texts which sound as if they had been composed in their own language, whereas the latter expect from the translation that it transmits part of the flair of the original language, as if it could be partly transparent. These differences are discussed in Venuti 1995 (I am grateful to Dominik Wujastyk who pointed out the book). The other element of the choice is the reader one has in mind. The first model presupposes a reader who knows Sanskrit fairly well and uses the translation only as an auxiliary, the second one vice versa.

A third element worth considering regards the translator themselves. They need an extreme command of English in order to translate in the second way. Moreover, they need to think of the duration of one’s translation. Each language rapidly evolves and translating in a very idiomatic way runs the risk of rendering the text non understandable for non-native speakers or speakers who will live in a not so distant future. As a non-native speaker of English, I have for instance had problems deciphering the English idioms at stake in a translation by Anand Venkatkrishnan (in Venkatkrishnan 2014, see the discussion here) and many readers of mine had asked me what “ones” in Edgerton 1929 could mean.

Anyway, even if one favours the latter type of translation, translating a Sanskrit text remains always a difficult task. In my opinion, this is due especially to the fact that contemporary readers lack almost completely the background assumptions which are needed in order to understand each instance of communication, including philosophical texts. For this reason, one might decide to strongly alter the text, in order to substitute the background assumptions with ones more familiar to the contemporary reader. This substitution may regard minor details, e.g., the substitution of “Devadatta” with “John Smith” as the placeholder for whoever a person, or the inclusion of pronouns, punctuation and other elements which can be deduced out of the context or of the śāstric usage. But these minor details are not likely to be enough, when it comes to philosophical positions which are not shared, such as the one about siddha and sādhya.

Thus, one needs more than a more reader-friendly translation. One might decide among at least three options:

1. Adding extensive comments in footnotes or endnotes (as it has been done in Preisendanz 1994)
2. Adding the same comments within the text in separate paragraphs, perhaps in smaller font size (as it has been done in Taber 2005)
3. Adding the same comments in an extended introduction (as it has been done in Bilimoria 1988)

The choice partly depends on one’s target readers. Philologists are more likely to appreciate the first solution, whereas the latter two are more likely to appeal to a public of more general readers, who might be more interested in the problem than in the text itself.

How do you traslate?
For another post on Sanskrit translations from a methodological point of view, see here.
Matthew Dasti raised a related point in this post, where he asks whether we should masculine pronouns to Sanskrit texts whenever they lack a subject and we are translating them in a language like English, which requires a subject.

Some preliminary work is surely needed. The following points immediately come to my mind:

1. Keep a list (to be constantly updated) of all technical terms and be sure that you discuss a unitary translation before completing your part.
2. Be sure you agree about the style of your translation. I, to begin with, use (round) parentheses for explanations and Sanskrit terms and [square] brackets for additions of words which are not found in the Sanskrit although they could have been there. E.g. nagaraṃ gacchāmi “I go to the city” (No need to put “I” in brackets, since it is already included in gacchāmi, same applies to “to the”). But kaiścid bhedo uktaḥ viṣayāntarāt “Some said that the difference [between Linguistic Communication and inference] is due to the fact that they have a different object” (since the author could have spelt out the elements which are different, but decided not to do so).
3. Be sure you agree about the purpose of your translation. As far as philosophical texts are concerned, for instance, I aim at being understandable while not camouflaging the style of the author. In other words, I would add several explanations in brackets if the text just presuposses things unknown to Western readers (e.g., that “Śyena” is the name of a malefic sacrifice), but I would not make the text sound as if it had been written yesterday in Austin (this is also one of the reasons, IMHO, for avoiding translations such as “inference to the best explanation” for arthāpatti).
4. Closely connected to the above is an agreement concerning one’s target reader. Kataoka 2011, for instance, clearly envisions a targer reader who knows Sanskrit and uses the translation as a guide through the text.
5. All translations of Sanskrit texts need an accurate introduction and/or footnotes and/or glosses and/or…. Be sure that you agree about how to use each of these tools (Should the introductory study explain all and make the following Sanskrit text understandable as in Freschi 2012? Should the introductory text show the philosophical relevance of the topic, while a close commentary follows each paragraph and explains it as in Taber 2005? and so on).
6. Just as a personal aside, I would also recommend to agree about a realistic timing. I find it very frustrating to keep a dead line, notwithstanding all, and then have to wait for moths for the others to be ready with their part.

What am I forgetting? Is there something else one should agree about before starting a joint project?

In what was possibly the first review of this book, John Taber (1983) felt nonetheless the need to point out that the chapters on Sabara should not be seen as faithful interpretations of Sabara’s thought. He is less strict in warning readers about the chapters on Kumarila, possibly because after the ones on Sabara, D’Sa seems to draw closer to texts and in this sense he does indeed look more like a “normal” scholar.

Recently, however, I re-read the section by D’Sa on the nature of sabda (pp. 116–122), since a student had based his thesis on it. This made me think that inexpert readers might still need some guidance through it:

1. varna does not mean “syllable”. k and u are two separate phonemes in the syllable ku
2. nitya does not mean “eternal”. The Mimamsa theory of communication does not need eternality at all. In fact, it just does not need temporality. All that is at stake is that sabda remains the same throughout each instance of its use, otherwise no communication could be possible. Accordingly, nitya means “fixed”. By the way, readers might remember that nitya does not originally mean “eternal”, although this became its most well-known meaning today. Minoru Hara has shown that nitya originally meant “own” (as sva). Within Mimamsa, it first (diachronically) and foremost (synchronically) means “fixed”, as when discussing the fixedness of nitya sacrifices.
3. sabda is defined as what is perceived by the ear, and not as what conveys a meaning. Accordingly, varnas are sabda. There is no space for a further sphota-like metaphysical entity called sabda. When Kumarila speaks of sabda as being indivisible, fixed and partless he is speaking of the only sabda he knows of, namely the varna.
4. Instead, what is different than the sabda-varna and manifests it are the single articulated sounds (dhvani) manifesting the immutable phonemes. All variations between the phoneme a as pronounced by me or by a native speaker of Sanskrit or English only depends on the dhvanis and not on the nature of the varna, which remains the same, as proved by the fact that I can nonetheless communicate with native speakers of English (or Sanskrit).

Do you want to correct some long lasting mistakes? Do it in the comment section

I just read Kiyotaka Yoshimizu’s Kumārila’s Reevaluation of the Sacrifice and the Veda from a Vedānta Perspective (in Mīmāṃsā and Vedānta, edited by Bronkhorst and Preisendanz, 2007). The paper elaborates on thematics close to the ones dealt with by Roque Mesquita (Die Idee der Erlösung bei Kumārilabhaṭṭa, WZKS 1994) and John Taber (Kumārila the Vedāntin?, in the same Mīmāṃsā and Vedānta) and adds to the debate Yoshimizu’s close knowledge of Kumārila in general and of his less studied works in particular. The article focuses in fact on the Ṭupṭīkā, Kumārila’s commentary on the last part of the Pūrva Mīmāṃsā sūtra, and compares it with the fragments of the Bṛhaṭṭīkā and with the Tantravārttika.

Kumārila is the chief exponent of the Bhāṭṭa school of Pūrva Mīmāṃsā and the Mīmāṃsā is mainly a school of Vedic exegesis. The Vedic sacrifices necessarily require someone responsible for their performance and responsibility is explicitly said to be individual. In other words, the Vedic injunctions enjoin specific individuals and not human beings in general. Thus, they require a plurality of subjects.

However, Yoshimizu shows how Kumārila accepts the notion of a paramātman ‘supreme Self’ in different passages of his works. paramātman can be used as a synonym of God, Īśvara, but is mostly used as a synonym of the all-encompassing brahman. The latter would contradict the plurality of subjects which is required by Mīmāṃsā.

Thus, we need to imagine that Kumārila’s paramātman does not entail monism. What else could it mean, then, to say that liberation is the “attainment of the supreme Self” (paramātmaprāpti, TV, quoted in fn. 6). Given that the paramātman seems to be in all authors who mention it a single entity, the TV claim seems to entail that everyone achieves the dignity of the single paramātman. How can this not contradict pluralism?

One might suggest that pluralism only exists in the saṃsāra, but could a pluralistic ontology be compatible with its monistic evolution, given that the paramātman is said to exist also along the saṃsāra? Would it make sense to think of living beings as leaving the proscenio of their plural world one after the other, in order to dissolve into the paramātman?

Alternatively, one should think of Kumārila’s claim as entailing an ontology akin to the one later known as Viśiṣṭādvaita Vedānta, i.e., only God exists independently, but human beings are his features (viśeṣa) and are, hence, not identical with him.

Can you think of other ways out?

P.S. Yoshimizu kindly informed me that he might elaborate further on the topic of the paramātman a new paper for the next World Sanskrit Conference in Bangkok.